BehaviorPileGroupUnderLateralLoads.pdf

7/27/2019 BehaviorPileGroupUnderLateralLoads.pdf

1/141

BEHAVIOUR OF PILE GROUPS UNDER LATERAL LOADS

A THESIS SUBMITTED TO

THE GRADUATE SCHOOL OF NATURAL AND APPLIED SCIENCESOF

MIDDLE EAST TECHNICAL UNIVERSITY

BY

ANIL ERCAN

IN PARTIAL FULFILLMENT OF THE REQUIREMENTSFOR

THE DEGREE OF MASTER OF SCIENCEIN

CIVIL ENGINEERING

APRIL 2010


2/141

Approval of the thesis:

BEHAVIUOR OF PILE GROUPS UNDER LATERALLOADS

submitted by ANIL ERCAN in partial fulfillment of the requirements for thedegree ofMaster of Science in Civil Engineering Department, Middle EastTechnical Universityby,

Prof. Dr. Canan zgen ____________________Dean, Graduate School ofNatural and Applied Sciences

Prof. Dr. Gney zcebe ____________________Head of Department, Civil Engineering

Prof Dr. Orhan Erol ____________________

Supervisor, Civil Engineering Dept., METU

Examining Committee Members:Prof. Dr. Erdal oka ____________________Civil Engineering Dept., METU

Prof. Dr. Orhan Erol ____________________Civil Engineering Dept., METU

Prof. Dr. Kemal nder etin ____________________Civil Engineering Dept., METU

Asst. Prof. Dr. Nejan Huvaj Sarhan ____________________Civil Engineering Dept., METU

Dr. zgr Kuruolu ____________________Yksel Proje

Date: 28.04.2010


3/141

iii

I hereby declare that all information in this document has been obtainedand presented in accordance with academic rules and ethical conduct. I

also declare that, as required by these rules and conduct, I have fully citedand referenced all material and results that are not original to this work.

Name, Last Name : ANIL ERCAN

Signature :


4/141

iv

ABSTRACT

BEHAVIOUR OF PILE GROUPS UNDER LATERAL

LOADS

Ercan, Anl

M.S., Department of Civil Engineering

Supervisor: Prof Dr. Orhan Erol

April 2010, 123 pages

To investigate the lateral load distribution of each pile in a pile group, the

bending moment distribution along the pile and the lateral group displacementswith respect to pile location in the group, pile spacing, pile diameter and soil

stiffness three dimensional finite element analysis were performed on 4x4 pile

groups in clay. Different Elatic Modulus values, pile spacings, pile diameters

and lateral load levels used in this study. In the analysis PLAXIS 3D

Foundation geotechnical finite element package was used. It is found that,

lateral load distribution among the piles was mainly a function of row location

in the group independent from pile spacing. For a given load the leading rowpiles carried the greatest load. However, the trailing row piles carried almost

the same loads. For a given load, bending moment values of the leading row

piles were greater than the trailing row piles. On the other hand, as the spacing

increased group displacements and individual pile loads decreased under the

same applied load. However, this behavior was seen more clearly in the first

and the second row piles. For the third and the fourth row piles, pile spacing


5/141

v

became a less significant factor affecting the load distribution. It is also found

that, pile diameter and soil stiffness are not significant factors on lateral load

distribution as row location and pile spacing.

Keywords: Piles; Pile Groups; Pile Spacing; Finite Element


6/141

vi

Z

KAZIK GRUPLARININ YANAL YKLER ALTINDAK

DAVRANII

Ercan, Anl

Yksek Lisans, naat Mhendislii Blm

Tez Yneticisi: Prof. Dr. Orhan Erol

Nisan 2010, 123 sayfa

Aratrma konusu kil ierisindeki 4x4 kare kazk gruplar iin, verilen yanal

ykler altnda kazklardaki yk, moment dalmlarnn ve grupdeplasmanlarnn kazklarn grup ierisindeki yerleimi, kazk ara mesafesi,

kazk ap ve zemin rijitliine bal olarak nasl etkilendii amacna yneliktir.

Analizler; farkl deformasyon modl deerleri, kazk aralklar, kazk aplar ve

yanal yk deerleri kullanlarak, sonlu elemanlar yntemi ile 3 boyutlu olarak

gerekletirilmitir. Analizlerde PLAXIS 3D Foundation sonlu elemanlar

program kullanlmtr. Yaplan almalar sonucunda, yanal yk dalmnn

byk oranda kazklarn grup ierisindeki yerleimine bal olduu

grlmtr. Kazklarn sra ierisindeki yerleiminin ise yk dalmn daha

az etkiledii gzlemlenmitir. Ayn yanal yk altnda ilk sra kazklarn

dierlerine orana daha fazla yk tad, fakat ikinci, nc ve drdnc sra

kazklarn tadklar yklerin ise birbirlerine ok yakn olduklar grlmtr.

Her bir kazk iin kazk boyunca moment dalmlar incelenmitir ve yk

dalmlaryla benzer ekilde ilk sra kazklarn eilme moment deerlerinin


7/141

vii

dier sra kazklara oranla daha yksek olduu grlmtr. Bunun yannda,

kazk aralklar arttka kazk gruplarnda meydana gelen deplasmanlarn ve

kazklarn tadklar yklerin azald gzlemlenmitir. Fakat bu davrann ilkiki sra kazk iin daha belirgin olduu sonucuna ulalmtr. Bunlarn yan

sra, kazk apndaki ve zemin deformasyon modlndeki deiimin, kazklar

aras yk dalmda daha az etkili olduu gzlemlenmitir.

Anahtar Kelimeler: Kazklar; Kazk Gruplar; Kazk Aral; Sonlu Elemanlar


8/141

viii

To My Family...


9/141

ix

ACKNOWLEDGEMENTS

I would like to express my special thanks to my dear supervisor, Prof. Dr.

Orhan Erol, for his brilliant ideas, endless support and guidance throughout this

study. I am grateful that, he did not only provide support about this study but

also shared his invaluable experience about life.

I would like to express my gratitude to my family; my lovely mother, Ayfer,

my powered father Ahmet, and my ingenious brother, Atakan.

Can Ozan Kurban deserves my grateful thanks for motivating, supporting and

helping me to complete this study as he does in all parts of my life.

It is with pleasure to express my deepest gratefulness to Yksel Proje managers

esspecially to my bosses Mr. Atilla Horoz and Dr. zgr Kuruolu for their

endless patience and tolerance throughout the research.

Sincere thanks to my friends for their precious friendship and continuous

support. Especially, the friends who look after me in bad times are invaluable.


10/141

x

TABLE OF CONTENTS

ABSTRACT ............................................................................................. iv

Z ........................................................................................................... vi

ACKNOWLEDGEMENTS ..................................................................... ix

TABLE OF CONTENTS .......................................................................... x

LIST OF FIGURES ................................................................................. xii

LIST OF TABLES .............................................................................. xviii

CHAPTERS

1.INTRODUCTION............................................................................................1

1.1 Theoretical Background....................................................................1

1.2 Research Objective............................................................................4

1.3 Scope of the Study............................................................................42.LTERATURE REVIEW.................................................................................5

2.1 Introduction.......................................................................................5

2.2 Full-Scale Tests.................................................................................5

2.3 Small-Scale Tests............................................................................11

2.3.1 Centrifuge Testing............................................................11

2.3.2 Other Model Tests............................................................17

2.4 Numerical Solutions........................................................................22

3.NUMERICAL MODELLING.......................................................................333.1 Introduction.....................................................................................33

3.2 Modeling Basics..............................................................................33

3.2.1 Definition of the Parametric Study and Analyzed PileGroups.......................................................................................33

3.2.2 Finite Element mesh and Boundary Conditions...............36

3.2.3 Intial Stress Conditions....................................................39


11/141

xi

3.3 Modeling Parameters.......................................................................39

4.DISCUSSION OF THE RESULTS...............................................................40

4.1 Introduction.....................................................................................404.2 Effect of Soil Stiffness to Load Distribution and Displacement ofGroup Piles...........................................................................................40

4.3 Effect of Pile Spacing to Load Distribution and Displacement ofGroup Piles...........................................................................................47

4.3.1 Load Distribution among Piles and Rows........................49

5.CONCLUSIONS............................................................................................56

REFERENCES..................................................................................................58

APPENDIX A...................................................................................................97


12/141

xii

LIST OF FIGURES

FIGURES

Figure 1.1 Typical p-y curve and resulting soil modulus .......................... 3

Figure 1.2 Basic formation of pile groups ................................................. 3

Figure 2.1 (a) Comparison of experimental and computed p-y curves fora single pile (b) experimental p-multipliers (fm) vs. depth (Brown et al.,

1988) .......................................................................................................... 6

Figure 2.2 Layout of single piles and pile groups (Rollins et al., 2006) ... 9

Figure 2.3 Average pile load-deflection curves for each rowin: (a) 3x3

pile group at 5.65D spacing; (b) 3x4 pile group at 4.4D pile spacing; (c)

3x5 pile group at 3.3D pile spacing compared with the single pile test

curve (Rollins et al., 2006) ................................................................. 9

Figure 2.4 Average pile load-deflection curves for each rowin: (a) 3x3

pile group at 5.65D spacing; (b) 3x4 pile group at 4.4D pile spacing; (c)

3x5 pile group at 3.3D pile spacing compared with the single pile test

curve (Rollins et al., 2006) ............................................................... 11

Figure 2.5 Sketch of the Model Setup ..................................................... 12

Figure 2.6 Average lateral load-pile displacement response (a) for pile

groups with 3 spacing (b) for pile groups with 5 spacing (= pile

width) (Ilyas et al.,2004) ............................................................. 17

Figure 2.7 Measured group efficiencies versus clear spacing for both in-

line and side-by-side configurations (Cox et al., 1984) .......................... 18

Figure 2.8 Comparison of group efficiencies for series and parallel

loading configurations (Rao et al., 1998) ................................................ 20


13/141

xiii

Figure 2.9 Surcharge loading adjacent to a pile group

(Bransby and Springman, 1995) .............................................................. 23

Figure 2.10 Geometry of the model (Bransby and Springman, 1995) .... 23

Figure 2.11 Layouts of single piles and pile groups

(Zhang, McVay and Lai, 1999) ............................................................... 26

Figure 2.12 Measured and predicted maximum bending moments in

individual piles of 4x3 pile group (a) in loose sand (b) in medium dense

sand (Zhang, McVay and Lai, 1999) .............................................. 27

Figure 2.13 Back-calculated p-multipliers for: (a) leading row; (b)

trailing row piles from this study and previous full-scale load tests along

with recommended design curves (Rollins et al., 2006) ......................... 29

Figure 2.14 Measured bending moment versus depth curves for each row

of 3x4 pile group at deflection of 25mm in comparison to curves

computed using GROUP and FLPIER with p-multipliers developed

during this study (Rollins et al., 2006) ................................................ 31

Figure 2.15 Soil and structure finite element simulation and plan view ofmodel simulation (Kahyaoglu et al., 2009) ............................................. 32

Figure 3.1 Mesh dimensions of the cross section of a typical 3D FE

model ....................................................................................................... 35

Figure 3.2 Mesh dimensions of a typical 3D FE model .......................... 35

Figure 3.3 Distribution of Nodes and Stress Points in a 15-node Wedge

Element (Plaxis 3D Foundation Manual, 2004) ...................................... 38

Figure 3.4 Typical pile group used in the parametric study .................... 40

Figure 3.5 Shematic illustration of lateral loading of piles (Active-pile-

loading) (Cubrinovski and Ishihara, 2007) ............................................. 44

Figure 3.6 The deflected forms of long and short piles subjected to

horizontal force at ground level; (a) a long pile with no head restraint; (b)

a long pile with a cap permitting no rotation of the head; (c) a short pile


14/141

xiv

with no head restraint; (d) a short pile with a cap permitting no rotation

of the head (Mohan, 1988) ...................................................................... 45

Figure 3.7 Shear and Bending Moment distribution along a fixed head

pile (Broms, 1964a) ................................................................................. 45

Figure 3.8 Curves for design of long piles under lateral load in cohesive

soil (Broms, 1964a) ................................................................................. 46

Figure 3.9 Interaction diagram use in determining the bending moment

capacity of a single pile with a diameter of 0.50m .................................. 49

Figure 4.1 Load Distribution with respect to row location for outer and

inner piles under same load applied (Pile Group with 3D pile spacing) . 57

Figure 4.2 Bending Moment vs. Depth Curves with respect to Row

Location for Outer and Inner Piles under same load (Pile Group with 3D

Pile Spacing, 8000kN total Load applied) .............................................. 58

Figure 4.3 Total Load vs. Group Displacement for different pile spacings

................................................................................................................. 60

Figure 4.4 Lateral Deflection Distribution along Length of Piles under8000kN .................................................................................................... 60

Figure 4.5 Total Load vs. Pile Load Curves with respect to Pile Spacing

for each Pile ............................................................................................. 65

Figure 4.6 Bending Moment Distribution Along Pile with respect to Pile

Spacing under 8000kN load applied ....................................................... 70

Figure 4.7 Pile Classification (DIN 4014) .............................................. 71

Figure 4.8 Pile Load Distribution in the group with 3D Pile Spacing and

8000kN load Applied .............................................................................. 73

Figure 4.9 Average Pile Load vs. Pile Load Coefficient with respect to

Pile Spacing ............................................................................................. 75

Figure 4.10 Interaction factors as a function of pile spacing .................. 78

Figure 4.11 Pile Load Coefficient values (Finite Element Analysis) ..... 79


15/141

xv

Figure 4.12 Pile Load Coefficient values (DIN4014) ............................. 80

Figure 4.13 Total Load vs Displacement (Pile Spacing is 3D) ............... 81

Figure 4.14 Total Load vs Pile Load (Pile Spacing is 3D) ..................... 83

Figure 4.15 Displacement curves for two differenet soil stiffness

(E=10Mpa, E=40MPa) ............................................................................ 85

Figure 4.16 Pile Load vs.Total Load Applied for two differenet soil

stiffness (E=10Mpa, E=40MPa) .............................................................. 89

Figure A.1 Load Distribution with respect to row location for outer and






Figure A.4 Bending Moment vs. Depth Curves with respect to Row


Pile Spacing, 1600kN total Load applied) ............................................ 100Figure A.5 Bending Moment vs. Depth Curves with respect to Row


Pile Spacing, 3200kN total Load applied) ............................................ 101

Figure A.6 Bending Moment vs. Depth Curves with respect to Row


Pile Spacing, 6400kN total Load applied) ............................................ 102

Figure A.7 Bending Moment Distribution Along Pile with respect to Pile

Spacing under 1600kN load applied ..................................................... 106






16/141

xvi

Figure A.10 Pile Load Distribution in the Group with 2D Pile Spacing

and 1600kN load Applied ...................................................................... 115














and 1600kN load Applied ...................................................................... 118Figure A.18 Pile Load Distribution in the Group with 4D Pile Spacing













17/141

xvii






and 11200kN load Applied .................................................................... 123


18/141

xviii

LIST OF TABLES

TABLES

Table 2.1 Relations Suggested p-multipliers for laterally loaded pile groups

(McVay et al., 1998)..14

Table 2.2 Comparison of Ultimate Lateral Resistance of Single Pile (Petra and

Pise, 2001).....22

Table 2.3 Comparison of Ultimate Lateral Resistance of 2x2 Pile Groups (Patra

and Pise, 2001)..22

Table 3.1 Variables of the parametric study..41

Table 3.2 Material Properties of Clay...42

Table 3.3 Material Properties of Pile and Pile Cap...43

Table 4.1 Load Distribution of Pile Groups with respect to Row Location andthe Location Within the Row for each Pile Group with different Pile Spacing

(D=0.50m, L=15m, E=40Mpa).52

Table 4.2 Load Distribution Coefficient of Individual Piles with respect to Row

Location and the Location Within the Row for each Pile Group with different

Pile Spacing (D=0.50m, L=15m, E=40Mpa)54

Table 4.3 Load Distribution Coefficient of Individual Piles.71


19/141

1

CHAPTER 1

1. INTRODUCTION

1.1 Theoretical Background

Many structures need deep foundations in order to utilize the bearing capacity

of deeper and stronger soil layers. Group piles are one particular type of deep

foundations most widely-used for high structures. In addition to the vertical

loads that must be carried by the piles, lateral loads may be present and must be

considered in design. These lateral loads can be caused by a variety of sources

such as earthquakes, high winds, wave action, ship impact, liquefaction, and

slope failure.

With respect to their use in practice, piles under lateral loads are termed active

piles or passive piles. An active pile is loaded principally at its top in

supporting a superstructure such as a brigde. However, a passive pile is loaded

principally along its length due to earth pressure, such as piles used as a

retaining wall in a moving slope.

The nature of the loading and the kind of soil around the pile, are major factors

in determining the response of an isolated single pile and the pile groups.

According to active loading at the pile head, four types can be identified: static

loading, cyclic loading, sustained loading and dynamic loading. Besides,


20/141

2

passive loadings can occur along the pile length due to moving soil, when a

pile is used as an anchor.

The curve in Figure 1.1 illustrates the case for a particular value of z where a

static loading is applied to a pile. Although this type of loading is encountered

seldom in practice, static curves are very useful since:

1- Analytical procedures can be used to develop expressions to correlate

with some portions of the curves,

2- The curves serve as a baseline for demonstrating the effects of other

types of loading, and

3- The curves can be used for sustained loading for some clays and sands.

(Reese and Impe; Single Piles and Pile Groups under Lateral Loading;

2001)

Piles are most widely used in groups as shown in Figure 1.2. The models that

are used for the group piles should reply to two problems:

1- The group efficiency of closely-spaced piles that are loaded laterally

2- Load distribution of individual piles in a group.

In the first case the forces are transmitted through the soil, however, in the

second case the forces are transmitted by the pile cap. In widely spaced pile

groups the pile-soil-pile interaction is inconsiderable and a solution is made in

order to reveal lateral load to each of the piles in the group.


21/141

3

Figure 1.1 Typical p-y curve and resulting soil modulus

Figure 1.2 Basic formation of pile groups


22/141

4

1.2 Research Objective

The aim of the study is to explore the effect of pile spacing, soil stiffness andthe load level on the load distribution of each pile in a pile group, bending

moment along the pile and the group displacements of the 4x4 pile groups in

clay. A numerical study on these factors using finite element analysis on

different cases of pile groups have been performed.

1.3 Scope of the Study

Following this introduction,

Chapter 2 presents an extensive literature review on the laterally loaded pile

groups. Full-scale and small-scale tests are illustrated first and then numerical

solutions are discussed.

Chapter 3 gives details of the numerical modeling. It defines the assumed pile

group arrangement and the pile and pile cap properties. Soil profile and soil

parameters are defined. Then, details regarding finite element model are given.

The chapter is concluded by presenting the material properties and construction

stages used in the analysis.

Chapter 4 includes the discussion of the results. For the static lateral loading,

effect of pile spacing on load distribution of each pile in a pile group is

discussed and FEM results are illustrated graphically

Chapter 5 presents major research findings and conclusions.


23/141

5

CHAPTER 2

2. LITERATURE REVIEW

2.1 Introduction

In literature, there are number of studies which deal with the laterally loaded

pile groups. These studies generally consist of two basic types, namely load

tests (full-scale and model tests) and numeraical solutions. Tests have been

performed since 1920s and provide a body of information concerning laterally

loaded pile groups. Full-scale tests are generally believed to provide the most

accurate results but, are rare due to the high costs. Therefore, many studies are

available concerning centrifuge and model testing. Evaluation of laterallyloaded pile groups has also been performed using numerical models. In many

studies, the results of the computer analyses that were performed using finite

element approach, were compared with the limited full-scale tests results. Most

of these computer analyses were performed in 3D Finite Elemet Programs,

rather than 2D modelling. In this chapter, results of previous research in these

areas will be discussed and summarized.

2.2 Full-Scale Tests

A study was carried out by Brown et al. (1988) in order to determine lateral

load behaviour of pile group in sand. In their study, a full-scale test was

conducted on a 3x3 pile group in medium sand underlain by very stiff clay.

The relative density of sand (Dr) is determined as 50%. Tests were performed


24/141

6

(a)

(b)

Figure 2.1 (a) Comparison of experimental and computed p-y curves for a

single pile (b) experimental p-multipliers (fm) vs. depth (Brown et al., 1988)


25/141

7

on nine closed-ended steel pipe piles that have 273mm of outer diameter and

9.27mm of wall thickness. The pile group was spaced at 3D on centers. Both

pile group and a single isolated pile were subjected to two-way cyclic lateralloading.

Brown et al. (1988) concluded that the pile group was observed to deflect

significantly more than the isolated single pile when loaded to similar average

load per pile. Moreover, the row position had an effect on the efficiency of the

individual piles. The front row (leading row) piles exhibited stiffer responsesthan the trailing rows (second and third row). However, no pattern was

observed of the pile position within a given row. The shadowing effect was

more considerable in sand compared to the clay as was mentioned in Brown et

al. (1987). However, when piles were under two-way cyclic loading, group

effects were still significant in sand, unlike the reduced significance of

shadowing with cyclic loading that was observed in clay.

The p-y curves were generated and typical p-y curves generated for single pile

are shown in Figure 2.1a. Moreover, p-multipliers concept was introduced and

this curve modified to group pile p-y curve for different depths and the results

are presented in Figure 2.1 b. As a result, Brown et al. (1988) suggested p-

multiplier values for the front, middle and back rows 0.8, 0.4 and 0.3

respectively.

Another study was conducted by Rollins et al. (1998) in order to investigate the

lateral load behaviour of pile groups in clay. Full-scale tests were performed on

a 3x3 pile group spaced at 2.82D with a pinned-head connection in soft to

medium-stiff clays overlaying dense sand. Moreover, in order to provide a


26/141

8

comparison, a single pile test was performed. For the tests, closed-end steel

pipe piles with an inner diameter of 0.305m and 9.5mm wall thickness were

chosen.

Rollins et al. (1998) concluded that, pile group deflection turned out to be more

than two times the single pile deflection for the same load level. In order to

provide a match between computed and measured results, the p-multipliers

method approach was used. As a result, Rollins et al. (1998) suggested p-

multiplier values for the front, middle and back rows 0.6, 0.38 and 0.43respectively.

Rollins et al. (2006) conducted another study to investigate group interaction

effects with respect to the pile spacing on laterally loaded pile groups. In order

to evaluate the behaviour, full-scale cyclic lateral load tests were performed on

3x5, 3x4 and 3x3 pile groups in stiff clay with 3.3D, 4.4D and 5.65D pile

spacing respectively, as shown in Figure 2.2. The soil profile generally consists

of stiff clay layers with sand layers that were in a medium compact density

state (Dr=60%), to a depth of 5m. These soils were underlain by sensitive clay,

silty clay and sand layers. Similar to the other studies, lateral load tests were

performed on single piles in order to provide comparison to the pile group test

results. For the tests, closed-end steel pipe piles with an outer diameter of

0.324m and 9.0mm wall thickness were chosen.


27/141

9

Figure 2.2 Layout of single piles and pile groups (Rollins et al., 2006)

(a) (b)

(c)

Figure 2.3 Average pile load-deflection curves for each rowin: (a) 3x3 pile

group at 5.65D spacing; (b) 3x4 pile group at 4.4D pile spacing; (c) 3x5 pile

group at 3.3D pile spacing compared with the single pile test curve

(Rollins et al., 2006)


28/141

10

Rollins et al. (2006) concluded that, lateral load resistance was a function of

pile spacing. While decreasing the pile spacing from 5.65D to 3.3D, group

interaction effects became progressively more important. Furthermore, as it canbe seen from Figure 2.3, the leading row (1st row) piles in the group carried the

greatest load, while the trailing row piles (second, third, fourth and fifth row

piles), carried smaller loads for the same displacement level. For these pile

groups driven in clay, row location within the group had more significant effect

on the lateral resistance than the location within a row. Figure 2.4 illustrates the

bending moment distribution for each row piles and it is concluded that, for the

same load level, the maximum moment in the trailing row piles were greaterthan in the leading row piles, due to the group interaction effects.

2.3 Small-Scale Tests

Small-scale tests generally consist of two types. These are the ones which

utilized a centrifuge and those that did not. The first section will discuss

centrifuge model tests and their results, and second section will discuss the

other small-scale tests results.

2.3.1 Centrifuge Testing

Centrifuge test is one of the most widely-used methods of conducting a model

test. The basic theory behind centrifuge modeling is similitude as well as

increased gravitational forces (Gerber, 2003). Ilyas et al. (2004) shows a

sketch of the model setup as shown in Figure 2.5. During the test, a model is

accelerated about an axis until the inertial forces reach to the gravitational

forces experienced by the prototype. The reduced cost of the test and the ability

to repeat tests with different parameters for comparison, are the major


29/141

11

advantages of this test. However, the difficulty in scaling is the major

disadvantage of this small-scale method (Gerber, 2003).

(a) (b)

(c)

Figure 2.4 Average pile load-deflection curves for each rowin: (a) 3x3 pile

group at 5.65D spacing; (b) 3x4 pile group at 4.4D pile spacing; (c) 3x5 pile

group at 3.3D pile spacing compared with the single pile test curve



30/141

12

Figure 2.5 Sketch of the Model Setup

(Ilyas et al., 2006)


31/141

13

A study was carried out by Barton (1984) in order to determine the response of

pile groups to lateral loading in the centrifuge. In tests, piles having diameters

that ranged from 9.5mm to 16mm that corresponds to prototype diameters from0.95m to 1.60m were used. During the tests, centrifugal acceleration varied

from 30g to 120g. Tests were conducted on both single piles and pile groups of

two, three and six piles with spacing of 2D, 4D and 8D.

In order to determine the group effects, interaction factors proposed by Poulos

(1971) were used in the analysis. The research was aimed mainly to evaluatethe accuracy of the elastic methods of analysis and determine the necessity of

non-linear analysis on model pile group response.

Barton (1984) concluded that the elastic method actually under-estimates pile

group interactions at a very close spacing. However, at a larger spacing this

method over-estimates the interaction factors. This shows that, soil non-

linearity has a significant affect on the strain field around a laterally loaded pile

even at small strains. Barton (1984), also concluded that the experimentally

derived factors for pairs of piles can be superimposed to give a good prediction

of the overall interaction factors for larger groups of piles.

Another study was carried out by McVay et al. (1998) in order to evaluate the

behaviour of laterally loaded pile groups in sand. Tests were conducted on 3x3

and 7x3 pile groups with 3D pile spacing. Moreover, single piles were tested in

order to provide comparison.


32/141

14

McVay et al. (1998) used p-multipliers method in order to predict the lateral

load behaviour of pile group. McVay et al. (1998) concluded that group

response and p-multiplier approach is independent of soil density, but mainly afunction of group geometry and row position. Table 2.1 shows the p-

multipliers suggested by McVay et al (1998) for each row.

Table 2.1 Relations Suggested p-multipliers for laterally loaded pile groups

(McVay et al., 1998)

Lateral load tests were performed in centrifuge in order to determine the group

effects by Remaud et al. (1998). Tests were conducted on a free-headed model

two-pile groups arranged at pile spacings of 2D and 6D. AU4G aluminium

hollow piles having 18mm outer diameter, 1.5mm wall thickness and 380mm

length were used during tests. These dimensions correspond to prototype piles

having 720mm outer diameter, 60mm wall thickness and 15.20m length. The

soil profile generally consists of Fontainebleau sand with a unit-weight of

16.3kN/m3.

Remaud et al. (1998) developed p-y curves using the bending moment profiles.

When the pile groups having different pile spacings were compared, group


33/141

15

effect was seen more clearly at the pile group having 2D spacing. The group

resistance decreased 20% for 2D spacing, however, pile group with 6D spacing

showed 5% resistance decrease. Moreover, at 2D spacing, the p-y curve on thetrailing pile was 50% of a single pile reaction but this value was reached to

93% for 6D spacing. On the other hand, the bending moments of the front row

pile was almost the same with the isolated single pile.

Much of the centrifuge model studies on laterally loaded pile groups were

conducted with soil profile consist of sand layers. Ilyas et al. (2004) is one ofthe relatively few studies on laterally loaded pile groups in clay. Centrifuge

model tests were performed both in normally consolidated and

overconsolidated kaolin clay. The piles were arranged symmetrically and the

groups consist of 2, 2x2, 2x3, 3x3, and 4x4 piles with 3 and 5 (= pile width)

spacing. In tests, hollow aluminium square tube piles having 12mm width and

260mm length were used. These model piles correspond to the prototype piles

having 840mm width and 18.20m length. All the tests were performed at 70g

on the National University Singapore Geotechnical Centrifuge.

Ilyas et al. (2004) concluded that, while increasing the number of piles in

group, the average lateral load per pile decreased. As Figure 2.6 illustrates

clearly, for piles installed in overconsolidated clay, the reduction of group

effect was less clear than for piles installed in normally consolidated clay.

Furthermore, for pile groups having 3 centre-centre pile spacing installed both

in normally consolidated clay and overconsolidated clay, group effect

decreased as the number of piles increased. However, for larger spasings (5),

group effect became recessive.


34/141

16

Ilyas et al. (2004) also concluded that, the shadowing effect was occured on

lead piles over trailin piles and this effect increased as the number of piles in

group increased. Thus, higher lateral loads were carried by the lead row piles.When the average load per pile was compared among the piles within a row,

the centre piles carried much less load and bending moment than the outer

piles.

2.3.2 Other Model Tests

A study was conducted by Cox et al. (1984) in order to determine the

behaviour of laterally loaded pile groups in very soft clay. Tests were

performed on both single piles and group piles to provide a comparison. The

pile groups consist of three and five piles with a clear spacing of 0.5D, 1D, 2D,

3D and 5D. Piles were arranged both in side-by-side and in-line configuration.

In side-by-side configuration, loading was perpendicular to the line of the piles.

However, in in-line configuration, loading was parallel to the line of piles. For

the tests, steel pipe piles having a wall-thickness of 0.71mm were chosen and

the soil was inorganic clay of high plasticity (PI = 40).

Cox et al. (1984) concluded that 2D or 3D clear pile spacing was enough to

produce resistance to match that of a single pile in side-by-side configuration.

For the in-line configuration, the load distribution depended on the pile group

horizontal displacement and the group efficiency decreased with the increase ofnumber of in-line piles from three to five. About 5D and 6D clear spacing was

enough to 100% group efficiency as illustrated in Figure 2.7..


35/141

17

(a)

(b)

Figure 2.6 Average lateral load-pile displacement response (a) for pile groups

with 3 spacing (b) for pile groups with 5 spacing (= pile width)

(Ilyas et al.,2004)


36/141

18

Figure 2.7 Measured group efficiencies versus clear spacing for both in-line

and side-by-side configurations (Cox et al., 1984)

Another study was conducted by Rao et al. (1998) in order to determine the

influence of rigidity on laterally loaded pile groups in marine clay. In tests,

marine clay deposits of India that had a PI of 30 were used. As Rao et al.

(1998) suggested that pile-fixity condition is closer to a free-headed

configuration, the piles were fixed to a pile cap which was a thin aluminum

plate. In tests, aluminum and mild steel pipe piles with different diameters and


37/141

19

embedment ratios (L/D) were used. 1x2, 2x3 and 1x4 pile groups were loaded

both in series and in parallel. In series loading, the groups were loaded parallel

to the pile line. However, in parallel loading, the groups were loadedperpendicular to the pile line.

In this study finite element method (FEM) analysis was also used for

comparison. The piles were defined as shear beams, and the pile cap or

aluminum plate was defined as a thin plate that connected the pile heads. The

results were compared as shown in Figure 2.8.

Rao et al. (1998) concluded that pile groups of short and rigid piles showed

greater resistance when loaded in parallel than in series, which means that

strength of the soil was more effective on the rigid pile deflection. However,

pile groups of long and flexible piles showed greater resistance when loaded in

series, which means that pile strength was more effective.

Patra and Pise (2001) conducted a study on ultimate lateral resistance of pile

groups in sand. In this study both single piles and group piles of 2x1, 3x1, 2x2,

and 3x2 configurations with 3D to 6D pile spacings were tested. Aluminum

alloy pipes having an outer diameter of 19mm and a wall thickness of 0.81mm

were used as model piles. Length to diameter ratios of the piles were equal to

12 and 38. The soil profile generally consisted of dry Ennore sand from

Chennai, India. In this study soil-pile friction angle was another variable. The

tests were repeated for = 20 and = 31.


38/141

20

Figure 2.8 Comparison of group efficiencies for series and parallel loading

configurations (Rao et al., 1998)

Approaches developed by Meyerhof et al. (1981), and Prasad and Chari (1999)

were used to compare the experimental results and observations of Patra and

Pise (2001). The predicted values of the ultimate lateral resistance for single

and group piles are presented in Table 2.2 and Table 2.3 with the values


39/141

21

observed by Meyerhof et al. (1981), and Prasad and Chari (1999). The results

observed by Patra and Pise (2001) were in close agrrement with the values

predicted by Meyerhof et al. (1981), and Prasad and Chari (1999). As shown inTable 2.2, Meyerhof et al. (1981) and Prasad and Chari (1999) underestimates

the single pile capacity. However, as shown in Table 2.3, Meyerhof et al.

(1981) overestimates the group capacity.

2.4 Numerical Solutions

A study was conducted by Bransby and Springman (1995) in order to evaluate

the short-term behaviour of group piles when subjected to lateral loading

occurred by deformation of a clay layer under an adjacent surcharge load as

shown in Figure 2.9, using three dimensional finite element analysis. The

objective of the analysis was to search on the pile-clay interaction behaviour.

The geometry modelled in this study was the same geometry modelled in

centrifuge tests performed by Bransby (1995) in order to provide a comparison.

As illustrated in Figure 2.10, the pile group consisted of two infinitely long

rows. Piles having 1.27m diameter and 19m length, were embedded in a 6m

layer of clay overlying dense sand. There was a 5m distance between the rows

and the piles were at a center-center spacing of 6.67m along the row. The pile

cap had a 9m of width and 1m of thickness and it was considered to be rigid.

Over the clay layer, there was a 1m layer of sand that applied a uniformsurcharge of 17kPa. However, the increasing uniform vertical surcharge load

was applied 1m away from the pile cap to the surface.


40/141

22

Table 2.2 Comparison of Ultimate Lateral Resistance of Single Pile

(Patra and Pise, 2001)

Table 2.3 Comparison of Ultimate Lateral Resistance of 2x2 Pile Groups

(Patra and Pise, 2001)


41/141

23

Figure 2.9 Surcharge loading adjacent to a pile group

(Bransby and Springman, 1995)

Figure 2.10 Geometry of the model (Bransby and Springman, 1995)


42/141


43/141


44/141

26

pile groups in stiff clay with 3.3D, 4.4D and 5.65D pile spacing respectively.

The soil profile generally consists of stiff clay layers with sand layers that were

in a medium compact density state (Dr=60%), to a depth of 5m. In computeranalyses same geometry and was modelled.

Figure 2.11 Layouts of single piles and pile groups

(Zhang, McVay and Lai, 1999)


45/141

27

(a)

(b)

Figure 2.12 Measured and predicted maximum bending moments in individual

piles of 4x3 pile group (a) in loose sand (b) in medium dense sand

(Zhang, McVay and Lai, 1999)


46/141


47/141

29

of the moving soil were determined by this parametric study. In order to verify

the accuracy of numerical analysis, existing experimental test results evaluating

the soil arching effect were re-examined.

Figure 2.13 Back-calculated p-multipliers for: (a) leading row; (b) trailing row

piles from this study and previous full-scale load tests along with

recommended design curves (Rollins et al., 2006)


48/141

30

For the analysis, the finite element analysis program PLAXIS 3D Foundation

was used. In Figure 2.15 both the plan view of model simulation and soil,

structure element simulation were shown. In models a wide range pile spacingranging from 2D to 8D was determined.

Kahyaoglu et al. (2009) concluded that results of the analysis performed using

PLAXIS 3D Foundation and the experiments were in close agreement. As the

pile spacing getting larger, the loads acting on the piles increased. However,

for the pile spacing larger than 8D, each pile behaved like an isolated singlepile without arching effect. The computer analysis results also showed that as

the pile spacing increased the residual load acting on the soil mass between

piles increased. In other words, for smaller pile spacings a small amount of

load acting on the soil between the piles was transferred to the piles. The soil

with higher internal friction angle developed stronger arching thus, more loads

transferred to the piles and less displacements occurred.


49/141

31

Figure 2.14 Measured bending moment versus depth curves for each row of

3x4 pile group at deflection of 25mm in comparison to curves computed using

GROUP and FLPIER with p-multipliers developed during this study



50/141

32

Figure 2.15 Soil and structure finite element simulation and plan view of model

simulation (Kahyaoglu et al., 2009)


51/141

33

CHAPTER 3

3. NUMERICAL MODELLING

3.1 Introduction

This study is focused on the assessment of the effects of pile spacing, pile

diameter and soil stiffness on lateral load distribution of each pile in a pile

group and bending moment distribution along the pile. A parametric study was

carried out for this purpose. Numerical analysis performed as a part of this

parametric study were carried out by Plaxis 3D Foundation geotechnical finite

element package which is specifically preferred for advanced analysis for piles

and pile-raft foundations. In the following paragraphs a short review of thisprogram is given.

Plaxis 3D Foundation program consists of four basic components, namely

Input, Calculation, Output and Curves. In the Input program the boundary

conditions, problem geometry with appropriate material properties are defined.

The problem geometry is the representation of a real three-dimensional

problem and it is defined by work-planes and boreholes. The model includes an

idealized soil profile, structural objects, construction stages and loading. The

model should be large enough so that the boundaries do not influence the

results. Boreholes are points in the geometry model that define the idealized

soil layers and the groundwater table at that point. Multiple boreholes are used

to define the variable soil profile of the project. During 3D mesh generation


52/141

34

soil layers are interpolated between the boreholes so that the boundaries

between the soil layers coincide with the boundaries of the elements. Work-

planes are horizontal planes with different y-coordinates that show the top-view of the model geometry. They are used to draw, activate and deactivate the

structural elements and loads. Each work-plane holds the same geometry lines

but vertical distance between them may vary. Within work-planes, points, lines

and clusters are used to describe a 2D geometry model.

After creating the 2D geometry model in a work-plane, a 2D is automaticallygenerated based on the composition of the clusters and lines in 2D geometry

model. 2D finite element mesh is composed of 6-nodes triangles. However, the

3D finite element mesh is the extension of 2D mesh into the third dimension

and it is generated after generating 2D mesh. The 2D mesh generation in the

program is fully automatic while 3D mesh generation is semi automatic. Mesh

dimensions should be appropriately defined, to prevent the effects of boundary

conditions. The 2D mesh should be constructed before proceeding to the 3D

mesh extension. Typical 2-D and 3-D meshes used in this study are presented

in Figure 3.1 and 3.2, respectively. To increase the accuracy, mesh width used

in the pile group was decreased. The mesh element size can be adjusted by

using a general mesh size varying from very coarse to very fine and also by

using line, cluster and point refinements. Very fines meshes should be avoided

in order to reduce the number of elements, thus to reduce the memory

consumption and calculation time. The program does not allow entering a new

structural element or a new soil cluster after the mesh is generated. If a new

element or cluster is added to the geometry model, the mesh generation should

be repeated with the new input.


53/141

35

Figure 3.1 Mesh dimensions of the cross section of a typical 3D FE model

Figure 3.2 Mesh dimensions of a typical 3D FE model


54/141

36

3D finite elemet mesh is composed of elements, nodes and stress points. While

generating the mesh, the geometry is divided into 15-node wedge elements. As

mentioned before, these elements are composed of the 6-node triangles in x-zdirection, as generated by 2D mesh generation. Moreover, 8-node quadrilateral

faces are generated in y-direction. The soil and the interfaces can be modelled

with different complexity levels. 6-node plate elements and 16-node interface

elements are used to model the soil-structure interaction.

The wedge elements that are used during mesh generation consist of 15 nodes.Figure 3.3 illustrates the distribution of nodes over the elements. Joining

elements are connected through their common nodes. During a finite element

analysis, displacement values are calculated at the nodes and a specific node

can be selected before calculation steps in order to generate the load-

displacement curves. On the contrary, stresses and strains are calculated at

individual stress points (Gaussian integration points) rather than at the nodes. A

15-node wedge element contains 6 stress points that shown in Figure 3.3.

However, stress and strain values at stress points are extrapolated to the nodes

for the output purposes.

At the bottom of the 3D finite element mesh, total fixities were used that

restrain the movements in both horizontal and vertical directions. For upper

part, 3D finite element mesh had no fixities. Besides, for right and left sides,

roller supports were used in order to restrain only the horizontal movements

and vertical displacements were left free.

After defining the model geometry and 3D mesh generation, initial stresses are

applied by using either K0-procedure or gravity loading. The calculation


55/141

37

procedure can be performed automatically or manually. The initial stresses in

the soil are affected by the weight of the soil and history of the soil formation.

Stress state is characterized by vertical and horizontal stresses. Initial verticalstress depends on the weight of the soil and pore pressures; whereas initial

horizontal stresses are related to the vertical stresses by the coefficient of

lateral earth pressure at rest. This relation is provided by the K0-procedure in

this study.

In this study, it is assumed that the water table is at the ground surface and clayformations are fully saturated; hence, initial stresses should be calculated in

terms of effective stresses. The relation between initial vertical and horizontal

stresses is given in Equation 3.1 and the coefficient of lateral earth pressure at

rest, K0, for normally consolidated soils can be calculated by Jacky (1944) s

formula as given in Equation 3.2.

' 'h 0 vK = (3.1)

'0K 1 sin = (3.2)

The construction stages are defined by activating or deactivating the structural

elements or soil clusters in the work-planes and a simulation of the

construction process can be achieved. A construction period can also be

specified for each construction stage but the soil material model should beselected as Hardening Soil Model.

The most important calculation type in Plaxis 3D Foundation is the staged

construction. In every calculation step, the material properties, geometry of the

model, loading condition and the ground water level can be redefined. During


56/141

38

the calculations in each construction step, a multiplier that controls the staged

construction process (Mstage) is increased from zero to the ultimate level that

is generally 1.0. The constructions that are not completed fully can be modeledby using this feature. (Plaxis 3D Foundation Manual, 2004)

This chapter is devoted to introduce the details of constitutive models, material

properties, and finite element modeling used in the performed parametric

study.

Figure 3.3 Distribution of Nodes and Stress Points in a 15-node Wedge

Element (Plaxis 3D Foundation Manual, 2004)


57/141

39

3.2 Modeling Basics

3.2.1 Definition of the Parametric Study and Analyzed PileGroups

This study was performed on a 4x4 pile groups with rows spaced at from 2D to

5D center-center in the direction of the loading as shown in Figure 3.4. As

illustrated in Figure 3.4 the piles were classified in the group according to their

row location and the location witihin the row. Leading row and trailing rows

were defined according to the loading direction. Moreover piles were defined

as outer and inner piles according to the location within the row. The purpose

of the analysis was to determine the individual pile behaviour within the group.

The load distribution of each pile and the bending moment distribution along

the pile at this particular pile arrangement was defined for clays.

The parametric study considers the variations in pile spacing, soil stiffness,

load level and pile diameter. The variables of the parametric study are listed in

Table 3.1. A total of 24 different combinations were studied using the variables

listed in Table 3.1.

For 3D numerical models, 4x4 pile group was used that shown in Figure 3.4.

The piles were fixed with a pile cap that has a thickness of 0.80m. The 15m

pile length was constant in all models. In order to determine the individual pile

behaviour with respect to the pile diameter, analysis had been performed with

two different pile diameters namely 0.80m and 0.50m. The pile spacing of 3D

was found to be sufficient for the explanation of load distribution with respect

to pile diameter. By increasing the applied force to observe the behavior of

0.80m pile diameter at 3D pile spacing in pile group, the loads acting on


58/141

40

LOADING

CENTER-CENTER

PILE SPACING

CENTER-CENTER

PILE SPACING

1st Row Piles

(Leading Row)

2nd Row Piles

(Trailing Row)

3rd Row Piles

(Trailing Row)

4th Row Piles

(Trailing Row)

Outer Piles

Inner Piles

Figure 3.4 Typical pile group used in the parametric study

individual piles were calculated. A wide range of pile spacing ranging from 2D

to 5D for the cases where the pile diameter was 0.50m had been investigated to

evaluate the influence of pile spacing. By increasing the applied force to

observe the behavior of 0.50m pile diameter, the loads acting on individual

piles were calculated for each pile spacing configuration.

In addition to pile diameter, pile spacing and lateral load level, other variable in

the parametric study was the soil stiffness. Analysis were performed on pile

group having 3D pile spacing for both soft and moderately stiff clay .


59/141

41

Table 3.1 Variables of the parametric study

D

=0.50m

E=40000kPa

Total Lateral Load

Applied to the System

(kN)

Center - Center Pile Spacing

2D 3D 4D 5D

1600

3200

6400

8000

9600

11200

E

=10000kPa

1600

3200

6400

8000

9600

11200

D

=0.80m

E=40000kPa

1600

3200

6400

8000

9600

11200


60/141

42

3.3 Modeling Parameters

3.3.1 Clay Parameters

Soil elements were modeled using Hardening Soil Model. Hardening Soil

model needs five input parameters, modulus of elasticity (E) and Poissons

ratio () to define elastic soil response; friction angle () and cohesion (c) to

define plastic response and angle of dilatancy (). The parametric study was

carried out for both soft clay and stiff clay layers. The parameters used in this

study are presented in Table 3.2. The dilatancy of the clay is not taken into

account in this study and full bond interface elements are used between soil

and piles.

3.3.2 Pile and Raft Parameters

Piles and pile cap were modeled using a linear elastic material model and the

corresponding material properties are given in Table 3.3.

Table 3.2 Material Ptoperties of Clay

Parameter Symbol Soft Clay Stiff Clay UnitType of Material Behaviour Drained Drained -

Unit Weight 19 19 kN/m3Poissons Ratio 0.30 0.30 -

Cohesion c 25 25 kN/m2

Internal Friction Angle 28 28

Elastic modulus E 10 000 40 000 kN/m2

Dilatancy Angle 0 0


61/141

43

Table 3.3 Material Ptoperties of Pile and Pile Cap

Parameter Symbol Pile Pile Cap UnitType of Material Behaviour Elastic Elastic -

Thickness t - 80 cm

Diameter d 50 - cm

Unit Weight 24 24 kN/m3

Elastic modulus E 28 500 000 28 500 000 kN/m2

Poisson's ratio 0.15 0.15 -

3.3.3 Loading Conditions

For the numerical simulation of the laterally loaded pile groups, 3D finite

element analysis had been performed at different load levels. These lateral

loads were determined depending on the lateral load capacity of a fixed-head

single pile in cohesive soil. The Broms method is the most widely-used

approach in practice. Thus, lateral load capacity of a single pile was calculated

using this method.

In the active pile loading case, the horizontal force at the pile causes the pile

deformation as shown in Figure 3.5. In this case, the mobilized earth pressure

provides an increasing resisting pressure with the increasing of the pile

deflection. On the other hand, plastic hinges can be occured on the critical

sections of the deflected piles due to the bending moment values larger than the

capacity. However, this behaviour is effected by the length of the pile and the


62/141

44

head fixity condition that is shown in Figure 3.6. A short pile rotates about a

point along its length. However, a long pile can not rotate and plastic hinges

and cracks occur at critical points.

In Broms approach it is assumed that ultimate resisting pressure is formed in

an idealized one layer soil profile. This resisting pressure depends on the pile

diameter (b) and undrained cohesion (c) of the soil. Broms (1964) eliminated

the soil resistance for the top 1.5b of the pile because of the lower resistnce in

that zone due to the pile deflection, below this level the pressure continuesconstantly as 9c as shown in Figure 3.7. In this study the piles were fixed

against rotation at their top by a pile cap and the lateral load capacity of fixed

head piles with a diameter of 0.50m and a length of 15m were calculated as

follows.

Figure 3.5 Shematic illustration of lateral loading of piles (Active-pile-loading)

(Cubrinovski and Ishihara, 2007)


63/141

45

Figure 3.6 The deflected forms of long and short piles subjected to horizontal

force at ground level; (a) a long pile with no head restraint; (b) a long pile with

a cap permitting no rotation of the head; (c) a short pile with no head restraint;

(d) a short pile with a cap permitting no rotation of the head (Mohan, 1988)

Figure 3.7 Shear and Bending Moment distribution along a fixed head pile

(Broms, 1964a)


64/141

46

Figure 3.8 Curves for design of long piles under lateral load in cohesive soil

(Broms, 1964a)

As mentioned before, short piles rotate without bending and reaches the

ultimate resisting pressure. Considering this behaviour, an equlibrium can be

written as follows and ultimate resisting force (P) can be calculated.

P = ( )bLbc 5.19 (3.3)

For a particular pile length the behaviour does not change. However, as the pile

length increses plastic hinges will be developed. At a particular pile length a

plastic hinge is developed at the pile head, and this length is termed as

intermediate length. For this condition, the point where the maximum


65/141

47

bending moment (bending moment capacity of the pile) occurs, can be

calculated from the following equations.

Mmax = ( ) yMfbP + 5.05.1 (3.4)

Mmax =225.2 gbc (3.5)

For the calculation of ultimate resisting force following two equations are also

needed.

L = gfb ++5.1 (3.6)

f =bc

P

9(3.7)

On the other hand, as the pile becomes longer, second plastic hinge is

developed at a critical point along the pile. This type of pile can be termed as

long pile. For this condition Mmax = My equilibrium can be wriiten and P

can be calculated from the following equation.

P =( )fb

My

+ 5.05.1

2(same as Equation 3.4) (3.8)

With these equtions, Broms presented a set of curves for solving the long pile

problem that shown in Figure 3.8. Entering the curve the value of My/cb3, P


66/141

48

value can be determined easily. However, first the pile length should be

determined. In order to determine pile length whether long or short, the

intermediate length should be calculated first, and this value can be found fromthe following equation with Equation 3.3.

P =( )bL

My

+ 75.05.0(3.9)

For the piles that were modeled in 3D finite element analysis, the lateral load

capacity was determined by following the procedure above. First, the bending

moment capacity of a bored pile having a diameter of 0.50m was calculated.

The interaction diagram shown in Figure 3.9 was used in order to determine

theis value. The axial load applied to the piles were assumed zero. Moreover,

the longitudinal reinforcement ratio was assumed about 4% as Turkish

Standards (TS500) suggested for coloumns. Thus, the moment capacity was

determined as 400kNm.

The parameters that were used in calculating the lateral load capacity of the

model piles are summarized below.

b = 0.50m

L = 15.00m

c = 100kPa

My = 400kNm


67/141

49

INTERACTION DIAGRAM

50 12 28

Material :

Concrete : C 20 cd = 133 g cm

Reinforcement : St 420 yd = 3652 g cm

Pile Diameter = 50 cm

Reinforcement Diameter = 28

Number of Reinforcement = 12

D = 50 cm Concrete Cover = 5 cm

Percentage of Reinforcement = % 3.76

-400.0

-300.0

-200.0

-100.0

0.0

100.0

200.0

300.0

400.0

500.0

600.0

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0

Moment ( ton-m )

AxialForce(ton)

AxialForce

(to

n)

Figure 3.9 Interaction diagram use in determining the bending moment

capacity of a single pile with a diameter of 0.50m


68/141

50

First, the pile length, where the pile goes from short pile mode of behaviour to

intermediate pile mode of behaviour, was calculated using the Equations 3.3

and 3.9. By solving these equations together, this value was found as 1.50mwhich is smaller than the model pile length. Thus, a second pile length, where

the pile goes from intermediate pile mode of behaviour to long pile mode of

behaviour, was calculated using the Equations 3.4, 3.5, 3.6 and 3.7. By solving

these equations together, this value was found as 4.00m which is also smaller

than the model pile length. Thus, the ultimate resistance force was determined

as 600kN using the curves for design of long piles suggested by Broms (1964).

In 3D finite element analysis, lateral loads were determined based on the lateral

load capacity of a single pile. Thus, the load applied to the pile groups varied

from 100kN/pile up to 700kN/pile.


69/141

51

CHAPTER 4

4. DISCUSSION OF THE RESULTS

This chapter presents of the results of the parametric study. A detailed

discussion on the role of important factors on lateral behavior of piles; such as

row location and pile location within a row, pile spacing, pile diameter, and

soil stiffness are also presented.

4.1 Effect of Row Location and Pile Location within a Row

The lateral load distribution among the individual piles and rows is a primary

concern, in order to understand group effects and various other behavioral

characteristics of pile groups. Here, the variation of the individual pile load

among the rows and within the row will be discussed.

Lateral load analysis for the pile groups were performed using computer

program Plaxis 3D Foundation and the load distribution was determined for

each row and for each pile within the row. The lateral load carried by the piles

was found to be a function of both row location and location within a row.

Table 4.1 presents the load distribution of pile groups with respect to row

location and pile location within the row for each pile group with different pile

spacings.


70/141

52

Table 4.1 Load Distribution of Pile Groups with respect to Row Location and the Location Within

with different Pile Spacing (D = 0.50m, L = 15m, E = 40Mpa)

1st ROW PILES

Total Load Applied to

the System (kN)Outer Pi les Inner Pi les Outer Pi les Inner Pi les Outer Pi les Inner Pi les Ou

1600 138 107 114 93 79 69

3200 370 247 278 227 181 161

6400 800 496 604 500 479 428

8000 1030 619 770 632 620 555

9600 x x x x 775 690 11200 x x x x x x

Individual Pile Load

(kN) in Group having 4D

Pile Spacing

(kN



Pile Spacing



Pile Spacing

2nd ROW PILES


the System (kN)Outer Pi les Inner Pi les Outer Pi les Inner Pi les Outer Pi les Inner Pi les Ou

1600 88 63 82 63 61 50

3200 188 141 177 139 131 110

6400 377 299 350 288 320 273

8000 466 388 434 368 395 340 9600 x x x x 475 412

11200 x x x x x x



Pile Spacing



Pile Spacing



Pile Spacing

(kN

52


71/141

53

3rd ROW PILES


the System (kN)Outer Pi les Inner Pi les Outer Pi les Inner Pi les Outer Pi les Inner Pi les O

1600 83 58 75 58 60 48

3200 160 117 163 125 122 98

6400 312 242 311 243 296 245

8000 385 308 378 305 359 300

9600 x x x x 423 354

11200 x x x x x x

Individual Pile Load(kN) in Group having 2D

Pile Spacing


Pile Spacing


Pile Spacing

(k

4th ROW PILES


the System (kN)Outer Pi les Inner Pi les Outer Pi les Inner Pi les Outer Pi les Inner Pi les O

1600 104 73 87 69 69 57

3200 172 128 169 137 143 120

6400 296 240 314 260 299 265

8000 354 295 370 320 360 324

9600 x x x x 426 383

11200 x x x x x x



Pile Spacing



Pile Spacing



Pile Spacing

(k

53


72/141

54

Table 4.2 Load Distribution Coefficient of Individual Piles with respect to Row Location and the L

different Pile Spacings (D = 0.50m, L = 15m, E = 40Mpa)

1st ROW PILES

Average Pile Load

(kN)Outer Pi les Inner Pi les Outer Pi les Inner Pi les Outer Pi les Inner Pi les Ou

100 1.38 1.07 1.14 0.93 0.79 0.69

200 1.85 1.24 1.39 1.14 0.91 0.81

400 2.00 1.24 1.51 1.25 1.20 1.07

500 2.06 1.24 1.54 1.26 1.24 1.11

600 x x x x 1.29 1.15 700 x x x x x x

Individual Pile Load /

Average Pile Load

(4D Pile Spacing)

IIndividual Pile Load /

Average Pile Load

(2D Pile Spacing)


Average Pile Load

(3D Pile Spacing )

2nd ROW PILES

Average Pile Load


100 0.88 0.63 0.82 0.63 0.61 0.50

200 0.94 0.71 0.89 0.70 0.66 0.55

400 0.94 0.75 0.88 0.72 0.80 0.68

500 0.93 0.78 0.87 0.74 0.79 0.68 600 x x x x 0.79 0.69

700 x x x x x x


Average Pile Load

(2D Pile Spacing )


Average Pile Load

(3D Pile Spacing )


Average Pile Load

(4D Pile Spacing)

I

54


73/141

55

3rd ROW PILES

Average Pile Load


100 0.83 0.58 0.75 0.58 0.60 0.48

200 0.80 0.59 0.82 0.63 0.61 0.49

400 0.78 0.61 0.78 0.61 0.74 0.61

500 0.77 0.62 0.76 0.61 0.72 0.60

600 x x x x 0.71 0.59

700 x x x x x x

Individual Pile Load /Average Pile Load

(2D Pile Spacing)

Individual Pile Load /Average Pil e Load

(3D Pile Spacing )

Individual Pile Load /Average Pile Load

(4D Pile Spacing )

I

4th ROW PILES

Average Pile Load


100 1.04 0.73 0.87 0.69 0.69 0.57

200 0.86 0.64 0.85 0.69 0.72 0.60

400 0.74 0.60 0.79 0.65 0.75 0.66

500 0.71 0.59 0.74 0.64 0.72 0.65

600 x x x x 0.71 0.64

700 x x x x x x


Average Pile Load

(2D Pile Spacing)


Average Pil e Load

(3D Pile Spacing )


Average Pile Load

(4D Pile Spacing )

I

55


74/141

56

As it was illustrated in Figure 3.4, the rows were defined as the leading row

and the trailing rows in the loading direction and piles were defined as outer

and inner piles according to their location within the row. Figure 4.1, illustratesthe summary of load distribution presented in Table 4.1 as a representative

case.

As expected based on the elastic theory, the piles located on the edges of a row

carry more load than the inner piles for an applied load. Moreover, the front

row piles (leading row piles) carried the greatest load while the second rowpiles carried succesively smaller loads under the same load applied. However,

the third and the fourt row piles carried about the same load. In fact, the fourth

row piles carried slightly higher loads than the third row piles.

Table 4.2 summarizes the load distribution among individual piles with a

coefficient of pile load for each pile in pile groups with different pile spacings.

These coefficients were calculated by dividing individual pile load calculated

from computer analysis with the assumed average pile load applied in the

analysis. Using Table 4.2 it is concluded that lateral load developed in outer

piles is about 1.25 times the load developed in inner piles. Moreover, although

this coefficient reaches to 1.5 or 2 for the leading row piles, coefficient of

trailing row piles decreases to 0.55 and 0.65 in some cases.


75/141

57

Pile Load vs. Total Load (3D - outer piles)

0

100

200

300

400

500

600

700

800

1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500

Total Load (kN)

PileLoad(kN)

1st Row

2nd Row

3rd Row

4th Row

Pile Load vs. Total Load (3D - inner piles)

0

100

200

300

400

500

600

700

1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500

Total Load (kN)

PileLoad(kN)

1st Row

2nd Row

3rd Row

4th Row

Figure 4.1 Load Distribution with respect to row location for outer and inner

piles under same load applied (Pile Group with 3D pile spacing)

4x4 Pile GroupPile Spacing = 3DPile Diameter = 0.50mPile Length = 15mOUTER PILES

4x4 Pile GroupPile Spacing = 3DPile Diameter = 0.50mPile Length = 15m

INNER PILES


76/141

58

Pile Depth vs. Moment (L=8000kN, 3D - outer piles)

-16.00

-15.00

-14.00

-13.00

-12.00

-11.00

-10.00

-9.00

-8.00

-7.00

-6.00

-5.00

-4.00

-3.00

-2.00

-1.00

0.00

-1000.00 -800.00 -600.00 -400.00 -200.00 0.00 200.00 400.00

Moment (kNm)

Depth(m)

1st row

2nd row

3rd row

4th row

Pile Depth vs. Moment (L=

-16.00

-15.00

-14.00

-13.00

-12.00

-11.00

-10.00

-9.00

-8.00

-7.00

-6.00

-5.00

-4.00

-3.00

-2.00

-1.00

0.00

-1000.00 -800.00 -600.00 -400.00 -

Moment (k

Depth(m)

Figure 4.2 Bending Moment vs. Depth Curves with respect to Row Location for Outer and Inner

Group with 3D Pile Spacing, 8000kN total Load applied)

4x4 Pile GroupPile Spacing = 3DPile Diameter = 0.50mPile Length = 15mApplied Load = 8000kNOUTER PILES

4x4 Pile GroupPile Spacing = 3DPile Diameter = 0.50mPile Length = 15mApplied Load = 8000kNINNER PILES

58


77/141

59

Bending moment versus depth curves are also shown with respect to row

location and pile location within the row in Figure 4.2 as a representative case.

Since all the piles were fixed to a pile cap, maximum moment occurred at pilehead as support moment. Parallel to the load distribution among piles, the piles

located on the edge of a row develop greater bending moment than the inner

piles under the same applied load. Lead row piles develop the maximum

bending moment while the trailing row piles develop somewhat smaller

moments under the same applied load. However, at greater depths lead row

piles develop less moment than the trailing row piles.

4.2 Effect of Pile Spacing

The alternative pile spacings used in this study have been introduced in Section

3.2.1. The load displacement curves of different pile spacings for pile groups

composed of 0.50m diameter and 15m long piles in stiff clay is presented in

Figure 4.3. Under the same lateral load applied, pile groups with 2D pile

spacing resulted in the largest lateral deflections, whereas pile groups with 5Dpile spacing resulted in the lowest lateral deflections. Pile groups with 3D and

4D pile spacings on the other hand, produced intermediate levels of deflection

as expected. Lateral deflection distribution along length of piles under 8000kN

total load applied is also presented in Figure 4.4. as an illustrative case. In

addition to the maximum displacement values of the pile groups with respect to

the pile spacing, in Figure 4.4, zero deflection point can be determined.


78/141

60

Group Displacement vs. Total Lo ad

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000

Total Load (kN)

GroupDisplacement(cm)

2D

3D

4D

5D

Figure 4.3 Total Load vs. Group Displacement for different pile spacings

Displacement vs. Depth Graph

-16.00

-14.00

-12.00

-10.00

-8.00

-6.00

-4.00

-2.00

0.00

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10

Displacement (m)

Depth(m)

2D

3D

4D

5D

Figure 4.4 Lateral Deflection Distribution along Length of Piles under 8000kN

4x4 Pile Group

Pile Diameter = 0.50mPile Len th= 15m

4x4 Pile GroupPile Diameter = 0.50mPile Length = 15mApplied Load = 8000kN


79/141

61

The lateral load capacity of a single pile was described in Section 3.3. Lateral

load capacity of a 0.50m diameter pile was calculated as 600kN. In Figure 4.4

the lateral deflection of an individual pile is presented under a total load of8000kN which means 500kN/pile. In other words, the results for this case is

presented under a load of 85% of its ultimate lateral load capacity. Using both

Figure 4.3 and 4.4, the increment of lateral deflection can be estimated for a

4x4 pile group depending on pile spacing. These figures reveal that lateral

deflection increased considerably as pile spacing decreased from 5D to 2D. In

this study, it is observed that for a 4x4 pile group under the same load, when

pile spacing decreases from 5D to 4D, maximum lateral deflection of the groupincreases about 33%. However, this increment of deflection is calculated larger

when pile spacing decreases from 4D to 3D and from 3D to 2D, namely 37.5%

and 64% respectively. Moreover, lateral deflection of individual piles in pile

groups with larger pile spacing reaches to zero at greater depths as expected. In

this study deflections become nearly zero between 7 m and 9 m depths from

the ground surface.

In this study, it is observed that pile spacing affects lateral load distribution in

pile groups significantly. In order to estimate the pile group behaviour, total

load versus pile load plots are shown in Figure 4.5 for each pile in the group.

Based on the variation of pile spacing, it can be concluded that as pile spacing

increases, pile load decreases. However, this type of behaviour can be seen

more clearly in the first and the second row piles. For the third and the fourth

row piles, pile spacing becomes a less significant factor affecting the load

distribution in a pile group.


80/141

62

Pile Load vs. Total Load (1st row - outer piles)

0

100

200

300

400

500

600

700

800

900

1000

1100

1000 2000 3000 4000 5000 6000 7000 8000 9000

Total Load (kN)

PileLoad(kN)

2D

3D

4D

5D

(a)

Pile Load vs. Total Load (1st row - inner piles)

0

Documents

BehaviorPileGroupUnderLateralLoads.pdf